CHIN LAYOUT_Layout 1 6/13/12 12:39 PM Page 48

R E C E N T A D VA N C E S I N
W I R E L E S S T E C H N O L O G I E S F O R S M A RT G R I D

OVERVIEW OF DEMAND MANAGEMENT IN

SMART GRID AND ENABLING WIRELESS
COMMUNICATION TECHNOLOGIES
ZIMING ZHU AND SANGARAPILLAI LAMBOTHARAN, LOUGHBOROUGH UNIVERSITY
WOON HAU CHIN AND ZHONG FAN, TOSHIBA RESEARCH EUROPE LTD.

ABSTRACT

Home area
network

The authors provide

an overview of
demand management with a particular focus on the
necessary enabling
wireless technologies. They review
various mechanisms
and algorithms for
the optimal demand
management in
smart grids using
these wireless technologies

48

There are significant challenges as well as

great opportunities for research at both policy
and technology levels on the efficient use of
energy. Most existing power generation and distribution systems are based on a century old
mechanism where power grids are managed by
vertically integrated utilities. Intelligent power
grids known as smart grids are required as the
demand for energy continues to grow and more
and more emphasis is being placed on the supply
of renewable energy. The main ingredient of
smart grids is the integration of information and
communication technology (ICT) into the grids
to monitor and regulate power generation and
demand. This article provides an overview of
demand management with a particular focus on
the necessary enabling wireless technologies.
Various mechanisms and algorithms for the optimal demand management in smart grids using
these wireless technologies are also reviewed.

INTRODUCTION
The electricity supply industry has been facing
significant challenges in terms of meeting the
projected demand for energy, environmental
issues, security, reliability and integration of
renewable energy. Currently, most of the power
grids are based on many decades old vertical
hierarchical infrastructures where the electric
power flows in one direction from the power
generator to the consumer side and grid monitoring information is handled only at the operation side. It is generally believed that a
fundamental evolution in electric power generation and supply system is needed to make the
grids more reliable, secure and efficient. This
can be achieved by enabling the future generation electricity network smarter and intelligent
by embedding bi-directional information and
communication architecture with power grids.
Various countries around the world such as
the US and several European countries have
already launched development projects on what

1536-1284/12/$25.00 2012 IEEE

is known as smart grids. For example, by 2030,

the US DoE Smart Grid R&D Program [1] aims
to achieve 20 percent reduction in the nations
peak energy demand, 100 percent availability to
serve all critical loads at any time and 20 percent
of the electricity capacity from distributed and
renewable energy sources. The European Smart
Grids Technology Platform has also set similar
targets with an emphasis on highly interconnected distribution networks and the integration of
renewable energy to meet the EU target on carbon emissions reduction by year 2020 [2].
There are various definitions for smart grid.
For example, the main highlight of the EU definition is that a smart grid is an electricity network that can intelligently integrate the behavior
and actions of all users to ensure sustainable,
economic, and secure electricity supply. The definition of US DoE states that a smart grid uses
digital technology to improve reliability, security,
and efficiency of the electricity system. Regardless of these different definitions, the main
ingredient of the smart grid is the application of
information and communication technology in
power grids. Figure 1 illustrates a possible overall smart grid architecture. It is a highly integrated and complex, yet flexible and reliable network
with various centralized and distributed energy
sources. As shown in Fig. 1, the power flow
direction is no longer just downhill from the
bulk power plants to consumers. Instead, dynamic flows can start from any generation sources
and end up anywhere in the grids. The energy
can be stored and released back to the grids
even at household level. The integration of ICT
enables not only the operation control center to
make informed decisions and optimize the energy flow, but it also provides opportunities for
consumers to participate in the energy demand
management to reduce the cost of their energy.
Demand management is the key to the operational efficiency and reliability of smart grid.
Facilitated by the two-way information flow and
various optimization mechanisms, operators benefit from real time dynamic load monitoring and
control while consumers benefit from optimized

IEEE Wireless Communications June 2012

CHIN LAYOUT_Layout 1 6/13/12 12:39 PM Page 49

Power line

Transmission

Bi-directional
communication link
Possible
communication link
Distribution
substation

Distributed
generation

Industrial
consumer

Transmission

Transmission
substation

Operation
control centre

Home
consumer

Bulk generation
Distributed
generation

Distributed
storage

Figure 1. Smart grid architecture.

use of energy. In this article, we provide a comprehensive overview of demand management in
smart grid with a focus on the enabling wireless
technologies. We first present survey on demand
management and associated research challenges.
We then discuss major features of various candidate wireless technologies required to facilitate
demand management. The features and enabling
mechanisms in terms of consumption scheduling,
real-time response and load balancing are also
discussed.

OVERVIEW OF DEMAND MANAGEMENT IN

SMART GRID
FEATURES
Demand management mainly consists of load
monitoring, analysis and response. In conventional power grids, the two sides of the electricity
demand and supply system are basically disconnected, as such demand management is performed exclusively by the utility operators using
mainly the raw data based local operation monitoring and state estimation. These approaches
have significant drawbacks in terms of high
response time (delay) and inaccuracy. The development of smart grid provides demand management with advanced features to enable many new
essential functions and applications as follows:

Bi-Directional Coordination In smart grid, demand

management is expected to be a combination of
centralized and distributed schemes. Monitoring
and control activities will not only be based at

IEEE Wireless Communications June 2012

the operation centers but can also be distributed

across the whole network. Every node at the
demand side of the network will be able to manage its own demand and consumption optimally
according to the current supply condition. These
activities will be acknowledged by the supply
side utility operators via effective bi-directional
information exchanges. Taking advantage of the
full visibility of the demand condition of the
grids, operators can alter their supply policies
such as price rates dynamically. Both sides of the
electricity market can participate in the demand
management and achieve a bi-directional coordination to fulfill customers requirements while
responding to the current circumstances of the
grid. This will reduce the management cost of
the grid operators and will potentially lead to a
win-win situation for the utility operators and
the consumers.

Data Gathering and Information Processing The

advanced instrumentation technology enabled by
real time sensing and data communication will
be the most important interface of the power
grids for monitoring the demand and supply. For
this purpose, the advanced metering infrastructures (AMI) have been proposed to gather and
convey real time raw measurement data.
Advanced signal processing techniques spanning
from data compression, data mining and optimizations will become important tools to extract
useful information from the raw data and to
generate appropriate demand and supply control
messages. Monitoring specific performance
parameters such as potential demand and local
back-up supply capability will enable grid opera-

49

CHIN LAYOUT_Layout 1 6/13/12 12:39 PM Page 50

The NANs and HANs

of AMI communications infrastructure
are particularly
suitable for wireless
deployment, largely
due to the ease and
low cost of adopting
wireless instead of
wired solutions.

tors to conduct more effective and accurate

demand management. The communication architecture will facilitate the data processing and
analysis to be performed either locally or distributively to reduce the workload of transmission and central controls.

Real-time and Online Processing Considering the

highly dynamic nature of the energy supply from,
for example renewable resources, in the electricity grids and the huge impact that can be caused
by possible control delays, it is important to handle the dynamics of the supply and demand in a
timely manner. As for bi-directional management in smart grid, effective communications is
of paramount importance. Modern communication and Internet technologies will ensure
prompt and transparent exchange of information
in the network. For example, after detection of a
potential outage, both the consumers and control authorities in the impacting area will be
notified immediately. Early actions can be taken
before further disturbances are spread. Local
area data processing and demand assessment is
subject to a minor delay of seconds so that associated control can respond effectively. These
activities are expected to be performed online
using various user interfaces. Every participant
of the activities will be responded and acknowledged transparently.
Proactiveness The success of the smart grid lies
in the full participation of the consumers. The
smart grid should enable everyone to have access
and participate in demand management. Importantly, the consumers should be given incentives
for participating proactively and coordinating
with the operators and other stake holders. To
achieve this, an efficient and transparent
exchange of information system facilitated by
advanced communication architecture and
attractive electricity consumption and price plans
are required. Proactive participation of the
demand side provides the operators not only the
opportunity to respond in real time to the supply
and demand, but also to predict the future
demand more accurately and devise appropriate
actions on the generation and supply of energy.

CHALLENGES
There are various perceived challenges spanning
from policy level to technology level including
social and behavioral aspects. The policy level
challenges include capital investment, enforcement rules on grid operators to provide considerable incentives to consumers, standardization
of electrical appliances and third party engagement of consumer raw data. The social and
behavioral aspects include trust and engagement
of consumers in the demand management. The
technological level challenge mainly spans the
integration of high quality and low delay twoway communication infrastructure with the
power grids. There are various state of the art
communication technologies available, however,
it is the choice of the most appropriate technology and the integration of all the components of
the smart grids that will form the important
challenge. To balance the supply and demand,
similar techniques as used in communication

50

networks for managing capacity of the network

and the resources can be used. For example,
optimization techniques using distributed strategies and game theory can be developed, as discussed in the succeeding sections.

WIRELESS TECHNOLOGIES
To facilitate demand management, suitable communications technologies must be chosen to
address various requirements in the different
parts of the AMI. The neighborhood area networks (NANs) and home area networks (HANs)
of AMI communications infrastructure are particularly suitable for wireless deployment, largely
due to the ease and low cost of adopting wireless
instead of wired solutions. The backhaul network connecting the AMI headend and the data
aggregation points (DAPs) can either be wireless
or wired. The AMI communication architecture
is illustrated in Fig. 2.
The link between the DAPs and consumers
requires NANs with a coverage in the range of
thousands of meters. Each DAP can connect to
hundreds of smart meters (SMs). As a result, a
key requirement of candidate wireless solutions
is coverage of wide area, which can also be
achieved through a mesh network architecture
or relay stations. Additionally, the wireless network must be able to provide a certain level of
reliability as well as low enough latency not only
to satisfy demand side management (DSM)
requirements but also to serve all other AMI
applications. According to communication
requirements from OpenSG [3], this translates
to a minimum reliability figure of 99.5 percent
and a latency requirement of less than 1 second,
which is a relatively relaxed figure as compared
to the commercial broadband requirements.
On the other hand, HANs which facilitate
energy management and planning within customer premises, require a relatively smaller coverage area. The requirements are also relatively
less stringent as there are less control messages
and information exchange between the smart
meter and smart appliances (and plug-in hybrid
electric vehicle (PHEV)). In general, the HAN
requires a minimum reliability figure of 99.5 percent and a latency requirement of less than 5
seconds [3].

NEIGHBORHOOD AREA NETWORKS

Candidate technologies for NAN have to provide coverage radius of over a thousand meter.
Reliability of communication channels between
the DAP and the smart meters dictates that the
spectrum used will have to be exclusive or interference free. Consequently, the most suitable
candidates need to be licensed or leased wireless
solutions. A comparison of the characteristics of
different NAN technologies can be found in
Table 1.

WiMAX Implementations of IEEEs 802.16

standard for metropolitan networks [4], commonly referred to as WiMAX (worldwide interoperability for microwave access), is a leading
candidate for providing connectivity between
DAPs and SMs. WiMAX is based on orthogonal division multplexing access (OFDMA),

IEEE Wireless Communications June 2012

CHIN LAYOUT_Layout 1 6/13/12 12:39 PM Page 51

which assigns slices of the frequency spectrum

to different users [5], avoiding interference
among the users and increasing the spectral
efficiency of the system. Although WiMAX is
not being widely adopted as a wireless broadband platform, it does not diminish its chance
of being a candidate as some utilities are
expected to set up dedicated DAPs. As a result,
WiMAX is more attractive in the sense that its
structure is much less sophisticated as compared to rival cellular standards such as 3GPP
Release 8 (commonly known as Long Term
Evolution (LTE)). Additionally, amendment j
to the standard added multihop relay capabilities [6], which can enlarge the coverage area
using low cost relay stations.

UMTS/LTE (2G/3G Cellular) Current cellular

technologies such as UMTS and LTE [7] also
provide attractive solutions for providing NAN
coverage. Relaying functionality had also been
incorporated in 3GPP Release 10 (commonly
known as LTE Advanced) [8], which will allow
extended coverage using relay/repeater stations.
However, the utilities have to be willing to overlay DAP-SM communications over existing communication infrastructure. Although the
advantage of overlaying is a lower setup cost
since the existing infrastructure can be used, the
utility operator will have to work with the
telecommunication operators to set up the network which can be contentious due to security
and privacy concerns.
IEEE 802.22 An alternative candidate to mainstream broadband wireless is the IEEE 802.22
wireless regional area network [9], which uses
white spaces in the television spectrum. The
IEEE 802.22 standard proposes to use cognitive
radio technologies to exploit unused spectrum in
the frequency spectrum allocated to television
broadcast. As the spectrum used is not dedicated, the latency in data transmission could be
higher as compared to other solutions mentioned earlier.

HOME AREA NETWORKS

Wireless solutions for HANs have a slightly different set of requirements, which are not as
stringent as those for NANs. In general, the
message arrival rate within a customer premise
is not as high as that between SMs and DAPs.
Additionally, the data volume is also much lower.
A comparison of various wireless candidate technologies for HAN is provided in Table 1.

WiFi IEEEs suite of standards for wireless

local area networks, IEEE 802.11 or WiFi, is the
most commonly deployed wireless standard within homes. As such, WiFi devices and chips are
relatively cheap, making it an attractive solution.
Amendment s of the standard also incorporates
mesh networking capability.
Zigbee Zigbee is one of the leading candidate
technologies for networking of devices in HANs.
The specification builds upon the IEEE 802.15.4
standard, and is tailored for mesh networking.
Zigbee also has various profiles to support different applications, such as Smart Energy. Zig-

IEEE Wireless Communications June 2012

DAP

DAP
AMI
headend

Home area
network

Backhaul network

Neighbourhood
area network (NAN)

Figure 2. Smart grid architecture.

bee Smart Energy 2.0 profile, which adds many

more features such as the support for PHEV,
will be ratified by the end of 2011.

Bluetooth The Bluetooth specification was

designed for personal area networks. The specification supports functions such as mesh networking. Furthermore, the specification ensures
less latency as compared to the two previously
mentioned standards through the use of a time
division multiple access (TDMA) like medium
access scheme. Both WiFi and Zigbee uses contention based carrier sense multiple access
(CSMA) which can result in large latency if
many devices are in operation.

DEVELOPMENTS OF
DEMAND MANAGEMENT APPROACHES AND
PROPOSED MECHANISMS
The success of smart grid lies in the design of
flexible and robust demand management techniques underpinned by the deployment of ICT
infrastructure mentioned earlier. Apart from
improving the legacy load control approaches,
the main contributions of recent research have
been in the demand side consumption scheduling, dynamic pricing and load balancing using
distributed energy resources (DER).

DEMAND SIDE CONSUMPTION OPTIMIZATION

Demand side consumption optimization is an
important feature to manage the (peak) demand
on the main grid and to maintain system reliability and stability. It has been an active research
topic for many years. For example, some
approaches in terms of peak clipping and flexible load shape shifting and related management
mechanisms have been outlined in [10]. However, it is the recent advancement of communication technology that has facilitated an entirely
new set of approaches and methods to perform
demand side management on a real time basis.
The operators could apply direct response and
control mechanisms through local or remote

51

CHIN LAYOUT_Layout 1 6/13/12 12:39 PM Page 52

As an analogy to the
design of hierarchical
topology based
Internet routing,
finding the most
suitable system
architecture of
consumption
management system
in smart grid is
an important
research topic.

52

Coverage

NAN

HAN

Technology

Range

Latency

Reliability

Cost & Ease of Deployment

WiMAX

30km

Low

High

Medium/Medium

UMTS/LTE

30km

Low

High

Medium/Low

802.22

30km

Medium

Medium

High/Medium

WiFi

200m

Medium-High

Low-Medium

Low

ZigBee

100m

Low-Medium

Medium

Low

Bluetooth

100m

Low

Medium

Low

Table 1. Comparison of candidate NAN and HAN wireless technologies.

control systems which directly control the energy

usage of different appliances in the customer
premises either coarsely by ON/OFF switching,
or by changing operational parameters such as
the temperature of hot water tank or heating
system. For indirect approaches, incentive based
management such as dynamic pricing or social
interaction can be adopted. The latter approach
provides more management flexibility and
enables proactive consumption optimization by
the distributed consumers, which might turn out
to be more cost-effective and efficient. It suits
better for managing real-time/daily consumption
and reducing peak loads. However, it has stricter
requirements in terms of metering technologies
(to support local analysis and computation) and
communication security. The centralized direct
management schemes are more suitable for
emergency response to prevent outages.
As an analogy to the design of hierarchical
topology based Internet routing, finding the
most suitable system architecture of consumption management system in smart grid is an
important research topic. For example, a three
step optimization methodology using a decision
tree structure was considered in [11]. The root
node acts as a global planner which tries to
achieve an overall control of the load profiles.
The root node can be the control centers at the
utility side. It decomposes the profiles in subparts and assigns them over its follower nodes in
the hierarchical structure. Follower nodes will
take the responsibility to plan its part of the consumption using similar optimization techniques
as the root node and will further decompose the
work into the leaf nodes. The leaf nodes are
directly linked to the controllers, e.g. smart
meters, located at the consumers terminals.
Communications between all the nodes are
essential to support the networked coordination.
As shown in Fig. 2, the AMI architecture
suits very well for the management structure.
The cost-effective and short-range Zigbee/Bluetooth based wireless sensor networks can be
deployed at the demand side to support
exchange of information between the leaf nodes
and the appliances. In the middle level which
represents NAN, the low latency and high reliability 3G/4G wireless solutions can be adopted.
High capacity wired technologies are suitable to
handle mass data flow at the top level between
DAP and AMI headend. However, the total

amount of communication workload is reduced

because of the hierarchical decomposition of
responsibility. Besides, the hierarchical decomposition supports scalability because the deployment of leaf nodes are reasonably independent
of other components of the tree.
The structure presented in [11] was designed
primarily for implementing global optimization
to regulate consumption. Indeed, mathematical
optimization is expected to be the heart of global consumption scheduling algorithms. For example, decentralized algorithms based on game
theory provides a flexible optimization framework for demand side management.
In [12], an energy scheduling framework
based on convex optimization technique was
proposed to schedule power consumption of
individual appliances to achieve the goal of peak
load reduction. The power requirements in
terms of the minimum standby power and maximum operating power of the appliances were
formulated into convex constraints. Both centralized and decentralized game theoretic frameworks were proposed. The authors showed that,
under the concave game settings, the participating users have the potential to quickly move
towards a unique point (Nash Equilibrium) at
which the consumption cost is optimized. Coordinated scheduling can be achieved using
HAN/NAN communication technologies mentioned earlier, and local computations are
required for conducting the optimization.
In practice, not all the appliances consumption requirements can be easily formulated into
convex forms. Some of the appliances may have
their own fixed consumption patterns to follow
which means once such an appliance is scheduled for operation, it has to work according to
its own power consumption profile until the end
of operation. Hence it is necessary to classify the
appliances into non-shiftable (fixed operation),
power-shiftable (schedule the operation power)
and time-shiftable (schedule the operation time)
groups.
We define non-shiftable appliances as those
that the scheduler cannot change the operation
time or power consumption profile. For example, a fridge is normally required to operate continuously for the whole day and a light bulb will
consume a certain level of power when it is
switched on. Power shiftable appliances are
defined as those whose power consumption pat-

IEEE Wireless Communications June 2012

Power

Power

CHIN LAYOUT_Layout 1 6/13/12 12:39 PM Page 53

Operation time

Operation time

Power

Power

(a)

Operation time

Operation time
(b)

Figure 3. Illustrative examples of power and time shiftable operations.

tern could be changed if required, as shown in

Fig. 3a. For example, a flexible battery charger.
Time-shiftable appliances are those that the consumers can tolerate postponing the operation.
For example, a washing machine or a dish washer. However, for these appliances, the power
may not be shiftable. For instance, the operation
of a washing machine could be postponed. However, when it is required to start, it will draw
power according to its own consumption pattern
as shown in Fig. 3b and this can not be changed.
The energy scheduling algorithm should consider
separate constraints for each class of appliances,
as explained in the following scheduling algorithm [13].
For a power-shiftable appliance a with a
standby power and a maximum operation
power , the scheduling requirement is written
as x a , where x a = [x a,1 , x a,2 , , x a,N ] T
denotes the scheduled power consumption vector over the day for the appliance and N determines the time resolution. For example, N = 24
for hourly scheduling and N = 1440 for minute
based scheduling. The parameter xa,t denotes the
power consumed by the appliance a at time t. In
addition, we will also require a constraint on the
total energy requirement, for example, N
t=1 xa,t
= la where la is the total energy requirement for
the operation of the appliance a.
A time-shiftable appliance b can have a preset power consumption pattern pb = [pb,1, pb,2,
, pb,N]T. We can only postpone the operation
of b, but the power consumption pattern should
remain the same, as shown in Fig. 3b. Hence the
scheduling result xb has to be exactly the same as
one of the cyclic shifts of the pattern pb. All possible shifts for the vector pb can be put together
in a matrix form as
p
b,1
p
Pb = b, 2

pb, N

pb, N

pb,1

pb, 4

pb, 2

pb, N 1

pb, 3

pb, 2

pb, 3
.

pb,1

(1)

We define a binary integer vector sb = [sb,1,

sb,2, , sb,N]T as the switch control for the timeshiftable appliance b. There is only one non-zero

IEEE Wireless Communications June 2012

Time-shiftable
appliances are those
that the consumers
can tolerate
postponing the
operation. For
example, a washing
machine or a dish
washer. However,
for these appliances,
the power may not
be shiftable.

element in the vector s b which is equal to one.

Now the schedule plan for a time-shiftable appliance b can be written as x b = P b s b , 1 T s b = 1.
The vector sb is an optimization parameter which
chooses the appropriate column of P b to optimize the energy consumption.
Using these definitions, we could formulate
the optimization problem as to minimize the
peak load L subject to the requirements of all
the appliances as follows:
min

L , x R+, s b Z+ N 1

s.t.

xa,t + xb,t L, t 1, 2,, N

a

b
N

x a , xa,t = la , a

(2)

t =1

xb = Pb s b , 1T s b = 1, b.
The above problem can be solved using mixed
integer linear programming (ILP) [13]. As an
example, we consider an hourly consumption
scheduling, i.e., N = 24, for a group of four
households. Every household is assumed to have
a similar set of appliances, however with different consumption requirements. Assume the total
daily requirement is 43 units. Without optimal
scheduling, it is likely that many appliances may
be operating simultaneously because users may
have similar consumption behaviors. The overlapping loads could generate undesirable peaks.
As illustrated in Fig. 4 (left), the loads are high
in the evening with a peak of 3.72kWh while
they remain low in the day time. Appliances
scheduling will reduce the peak demand through
coordination between households and adjusting
the operation time and the power of shiftable
appliances.
Figure 4 (right) depicts the optimal hourly
consumption scheduling result using the ILP
optimization in Eq. 2. The optimization schedules appliances to minimize the peak load L
over t = [1, 2, , 24]. As seen, there is no significant peak and the load allocation is fairly balanced. This is because the shiftable
consumptions are reallocated optimally through-

53

CHIN LAYOUT_Layout 1 6/13/12 12:39 PM Page 54

2.5
Scheduled load

Load without scheduling

Power consumption (kwh)

Power consumption (kwh)

3.5
3
2.5
2
1.5
1

1.5

0.5
0.5
0
0

10
15
Time (hour)

20

25

0
0

10 12 14
Time (hour)

16

18

20

22

2425

Figure 4. Hourly load over the day. Left: without scheduling, right [13]: after scheduling.

out the day. The optimized hourly peak load L is

2.14kWh and the peak to average load ratio is
2.14/(43/24) = 1.19, which means the peak load
is just 19 percent higher than the daily average.
In contrast, without scheduling, the peak to
average load ratio is 2.07. The optimization is
able to control the peak load at a reasonable
level. With a dynamic price plan, the ILP formulation can also be applied to incentive based
scheduling to minimize the energy cost which
will in turn attempt to balance the consumption
loads.
The scheduling of appliances may introduce
discomfort to the consumers mainly due to possible delays introduced by shifting the operation
of the appliances. A successful scheduling should
therefore ensure that the appliances are scheduled according to certain user preference. User
preference can be formulated into the optimization problem using additional constraints. Also,
considering the discomfort as a cost of inconvenience, this can be factored into the overall optimization cost. However, the reduction of user
discomfort is expected to increase the peak load
and probably the economic cost. Therefore, the
best strategy is to provide incentives to the consumers and encourage them to participate in the
demand management. A robust load balancing
algorithm should consider various pricing profiles by multiple utility providers and quantify
consumer behavior using numerical means to
optimize both the peak loads and incentive to
consumers.

DYNAMIC PRICING
Generally, issuing dynamic pricing policies as
user incentive is the most effective way to
achieve indirect demand management for the
grid side operators and controllers. Dynamic
pricing mainly consists of time-of-use (ToU)
pricing, critical peak pricing (CPP), time block
based pricing (TBBP), and real-time pricing
(RTP), as listed in Table 2.
ToU and CPP rates have already been included in many utility contracts/tariffs in the current
electricity markets for load control purposes. In
order to account for the dynamic demand in
smart grid, pricing policies should be updated

54

frequently. It is believed that a combination of

time block based pricing and ToU/CPP rates
could be a possible solution for the early stage
of smart grid development with limited ICT
deployments. Some economically driven consumption optimization algorithms, such as that
in [12], consider this kind of pricing schemes.
However, the performance of such pricing
mechanism will highly depend on the accuracy of
the demand estimation/prediction and risk
assessment. Self-learning algorithms can be used
for demand prediction. Various risk control
mechanisms used in business research can also
be adopted in the design of pricing policies.
RTP is believed to be robust in terms of
responding to the dynamics even when there is
unpredictable energy demand in the grid. The
main challenge of implementing RTP is the
expectation of a high quality communication
infrastructure for real-time monitoring purposes.
Latency will be the primary concern in choosing
the communications solution for RTP. In order
to support continuous and mass flow of data, the
throughput of the communication network
should also be very high. Finally, the power consumption of the communication infrastructure
itself has to be managed optimally.

DISTRIBUTED ENERGY RESOURCE (DER)

MANAGEMENT
In addition to balancing the supply and demand,
the smart grids when integrated with the distributed energy sources will enable the consumers to choose different type of energy sources
and suppliers as well as to optimally use and sell
back the locally generated and stored energy.
The optimal use of available energy at different
times can help reducing the dependency on the
central supply. The grid side utilities can balance
the load by choosing the supply from different
generation systems especially at peak demand
periods. For example, the work in [14] discusses
how various energy supplies can be aggregated
and dispatched. The idea of load-based services
can bring true benefit for the access of varying
energy generation from green resources (such as
wind and solar) and smart charging of PHEVs.

IEEE Wireless Communications June 2012

CHIN LAYOUT_Layout 1 6/13/12 12:39 PM Page 55

Pricing policy

Characteristics

Cost & ease of deployment

Time-of-use pricing

One-off issuing rates depending on the time of use Limited performance for dynamic demand control

User behavior and load estimation required

Low ICT requirements

Critical peak pricing

One-off issuing rates depending on particular events Critical rate

for pre-defined peak times (or loads) Limited performance for
dynamic demand control

Hard to define critical events Low ICT

requirements

Time block based

pricing

Monthly/weekly/daily updating rates ToU rates or load-sensitive

rates Enhanced performance for dynamic demand control

Load prediction and risk control required

Non-real-time ICT required

Real-time pricing

(Near) real-time updating rates Advanced performance for

dynamic demand control

Advanced real-time ICT required High

communications power consumption

Table 2. Dynamic pricing schemes.

Residential DER management based on instantaneous supply conditions (both from the central
energy source and the distributed energy sources)
is also an important research topic. For example,
various households in a neighborhood area can
share locally generated power and draw power
from the central energy source only when it is
required. The authors of [15] developed a decision-support algorithm using particle swarm optimization (PSO) to support this kind of schemes.

CONCLUSIONS
The main ingredient of smart grids is the integration of ICT into the grids to monitor and regulate energy generation and demand. The smart
meters and sensors will be deployed in various
parts of the grid, starting from the generation,
through distribution, and all the way to the
household level. These will be interconnected
through both wired and wireless connections.
Due to the consideration of cost effective implementation, wireless solutions are preferred at
the NAN and HAN levels and wired connections
could be used for backhaul networks. In addition
to the integration of renewable energy, one of
the main goals of the smart grids is to perform
demand management to reduce peak loads. This
requires acquision of real time data from various
points in the grid and optimization of the power
supply and demand. In order for the demand
management to be successful, consumers should
be given adequate incentives for full participation. This article covered various candidate communication technologies and mathematical
optimization algorithms to enable demand management. As communication is an underpinning
technology for the success of smart grid, we
envisage that smart grids will be an exciting
research area for communication engineers for
many years to come.

ACKNOWLEDGEMENTS
The authors would like to express sincere thanks
to Toshiba Research Europe Limited for funding this work.

REFERENCES
[1] US DoE, Smart Grid Research & Development: MultiYear Program Plan(MYPP) 20102014, Available
http://www.oe.energy.gov.

IEEE Wireless Communications June 2012

[2] European Commission, European Smart Grids Technology Platform Vision and Strategy for Europes Electricity Networks of the Future, Directorate-General for
Research Sustainable Energy Systems, 2006, available http://www.smartgrids.eu.
[3] OpenSG, SG Network Systems Requirement Specification v5.0 (draft), Aug. 2011, available
http://osgug.ucaiug.org.
[4] IEEE Standard for Local and Metropolitan Area Networks Part 16: Air Interface for Broadband Wireless
Access Systems, IEEE Std 802.16-2009 (Revision of
IEEE Std. 802.16-2004), May 2009, pp. C12004.
[5] H. Zhu and J. Wang, Chunk-based Resource Allocation
in OFDMA Systems Part I: Chunk Allocation, IEEE
Trans. Commun., vol. 57, Sept. 2009, pp. 273444.
[6] IEEE Standard for Local and Metropolitan Area Networks Part 16: Air Interface for Broadband Wireless
Access Systems Amendment 1: Multiple Relay Specification, IEEE Std. 802.16j-2009 (Amendment to IEEE Std.
802.16-2009), pp. C1290, Dec. 2009.
[7] 3GPP TSG RAN, Evolved Universal Terrestrial Radio
Access (E-UTRA); Physical Channels and Modulation
(Release 8), TS 36.211, V8.9.0, Dec. 2009.
[8] 3GPP TSG RAN, Evolved Universal Terrestrial Radio
Access (E-UTRA); Physical Channels and Modulation
(Release 10), TS 36.211, V10.4.0, Dec. 2011.
[9] IEEE Standard for Information Technology Telecommunications and Information Exchange Between
Systems Wireless Regional Area Networks (WRAN)Specific requirements Part 22: Cognitive Wireless RAN
Medium Access Control (MAC) and Physical Layer (PHY)
Specifications: Policies and Procedures for Operation in
the TV Bands, IEEE Std. 802.22-2011, Jan. 2011, pp.
1680.
[10] C. Gellings and W. Smith, Integrating Demand-Side
Management into Utility Planning, Proc. IEEE, vol. 77,
no. 6, 1989, pp. 90818.
[11] V. Bakker et al., Demand Side Load Management
Using A Three Step Optimization Methodology, 1st
IEEE Intl. Conf. Smart Grid Communications (SmartGridComm), Gaithersburg, Maryland, USA, Oct. 2010,
pp. 43136.
[12] A. Mohsenian-Rad et al., Autonomous Demand-Side
Management based on Game-Theoretic Energy Consumption Scheduling for the Future Smart Grid, IEEE
Trans. Smart Grid, vol. 1, no. 3, Dec. 2010, pp. 32031.
[13] Z. Zhu et al., An Integer Linear Programming based
Optimization for Home Demand-Side Management in
Smart Grid, IEEE PES ISGT, Washington DC, USA, Jan.
2012.
[14] A. Brooks et al., Demand Dispatch, IEEE Power and
Energy Mag., vol. 8, no. 7, pp. 2029, Jun. 2010.
[15] M. Pedrasa, T. Spooner, and I. MacGill, Coordinated
Scheduling of Residential Distributed Energy Resources
to Optimize Smart Home Energy Services, IEEE Trans.
Smart Grid, vol. 1, no. 2, Sept. 2010, pp. 13443.

BIOGRAPHIES
Z IMING Z HU received the M.Sc. degree in Communication
Networks and Signal Processing with Distinction from Bristol University, U.K., in October 2010. He is currently a Ph.D.
student within the Advanced Signal Processing Group
(ASPG) at Loughborough University, U.K. His research inter-

55

CHIN LAYOUT_Layout 1 6/13/12 12:39 PM Page 56

ests include resource management and optimization for

wireless networks and smart grids.
SANGARAPILLAI LAMBOTHARAN received the Ph.D. degree in signal processing from Imperial College London, U.K., in
1997. He was with Imperial College until 1999 as a Postdoctoral Research Associate. In 1996, he was a Visiting Scientist with the Engineering and Theory Center, Cornell
University, Ithaca, NY. From 1999 to 2002, he was with the
Motorola Applied Research Group, U.K., as a Research
Engineer, and investigated various projects, including physical-link layer modeling and performance characterization
of GPRS, EGPRS, and UTRAN. From 2002 to 2007, he was
with the Kings College London, London, U.K., and Cardiff
University, Wales, U.K., as a Lecturer and Senior Lecturer,
respectively. In September 2007, he joined the Advanced
Signal Processing Group, Loughborough University, Leicestershire, U.K., as a Reader, and was promoted to Professor
of digital communications in September 2011. He has published more than 100 conference and journal articles in
these areas. He serves as an Associate Editor for the
EURASIP Journal on Wireless Communications and Networking. His current research interests include multipleinput-multiple-output, wireless relay networks, cognitive
radio networks, and smart grids.

56

W OON H AU C HIN received the Ph.D. degree in electrical

engineering from Imperial College London, U.K., in 2004.
From 2000 to 2008, he was with the Institute for Infocomm Research, Singapore, where he was involved in the
standardization of IEEE 802.11n and 3GPP LTE. He was
also an Adjunct Assistant Professor at the National University of Singapore from 2005 to 2008. Since 2008, he
has been with Toshibas Telecommunications Research
Laboratory in Bristol, U.K., where he is a Principal
Research Engineer leading a research team and projects
on Smart Home technologies. His current research interests include cooperative/relay, energy efficient, and M2M
communications.
ZHONG FAN is a Chief Research Fellow with Toshiba Research
Europe in Bristol, UK. Prior to joining Toshiba, he worked
as a Research Fellow at Cambridge University, a Lecturer at
Birmingham University and a Researcher at Marconi Labs
Cambridge. He was also awarded a BT Short-Term Fellowship to work at BT Labs. He received his BS and MS degrees
in Electronic Engineering from Tsinghua University, China
and his Ph.D. degree in Telecommunication Networks from
Durham University, UK. His research interests are wireless
networks, IP networks, M2M, and smart grid communications.

IEEE Wireless Communications June 2012